Re^3: A list assignment gotcha (Updated)

[LanX]

Hi

> A bit trickier than I originally thought.

I was about to update that one needs :lvalue and returning a tied scalar then.

Surprised you made it without tie , my last use case must have been different than. :)

update

> You can use [3, 2, 1] on the RHS, as well.

Nice! :)

> on both the listifier and constructor.

Genius!!! that's how you avoided using Tie::Scalar! :)

[jcb]

Great work by both of you, and the possible syntax tricks nicely show TIMTOWTDI, but the original motive for trying LIST += LIST was a hoped-for performance gain from the interpreter completing the whole set in one opcode, instead of add...add...add....

I suspect that overloaded objects like this will cause problems if used in inner loops. :-) Neat tricks, though...

[LanX]

> will cause problems if used in inner loops

performance won't be great unless implemented in XS.

But that's the way I would implement new features in Perl.

• a nice syntactic sugar which does the job in pure Perl and stays fully backwards compatible.
• speed it up in XS for newer versions
• if it becomes popular, consider including it into CORE.

[jcb]

Even with an XS implementation, you will still have the dispatch overhead for the overloaded operator. Eventually including it in core perl would allow the parser to resolve the dispatch at compile-time, but until then I wonder how long the list needs to be to make up that difference?